International

Journal of the Sociology of Language. DOI 10.1515/ijsl-2016-0034

Language maintenance and shift under pressure:

Three generations of the Turkish immigrant community in the
Netherlands

Yeşim Sevinç
University of Oslo, Center for Multilingualism in Society across the Lifespan (MultiLing)

To cite this article: Sevinç, Y. (2016). Language maintenance and shift under pressure: Three
generations of the Turkish immigrant community in the Netherlands. International Journal of the
Sociology of Language, 242, 81-117. DOI:10.1515/ijsl-2016-0034

Abstract
Drawing on questionnaire and interview data, this study explores the process of language
maintenance and shift across three generations of Turkish immigrants in the Netherlands. It
compares three generations of Turkish-Dutch bilinguals by examining age and place of
language learning, self-rated language proficiency, and language choices in six domains
(home, school, work, friends, media and leisure time activities, and cognitive activities).
Furthermore, it investigates bilinguals’ experiences, motivations for learning languages and
attitudes towards bilingualism. Findings suggest that following the typical pattern of
language shift described by Mario Saltarelli and Susan Gonzo in 1977, language history, self-
rated language proficiency and current language practices of third-generation children differ
from those of first- and second-generation bilinguals. Consequently, possible language shift
among third-generation bilinguals causes socioemotional pressure about maintaining the
Turkish language, triggering intergenerational tensions in Turkish immigrant families. At the
same time, the perceived need to shift to Dutch for social and economic reasons causes
immigrant children to experience tensions and ambiguities in the linguistic connections
between the family and other social domains (e.g. school, friendship). The findings evidence
that the Turkish immigrant community in the Netherlands may no longer be as linguistically
homogeneous as once observed. The dissolution of homogeneity can be a sign of social
change in which maintaining the Turkish language has become a challenge, whereas
speaking Dutch is a necessity of life in the Netherlands.
Keywords: language shift, three generations, Turkish immigrant community in the
Netherlands, language tension, socioemotional pressure.

1
1 Introduction
Immigrant communities can experience tension between language maintenance and
language shift, influenced by countless imperceptible linguistic choices. Language
maintenance occurs when immigrants’ heritage language (HL)1 continues to be used over
successive generations (Fishman 1972). Conversely, language shift occurs when immigrants
progressively replace their HL with the language of the socially or economically dominant
group (Fishman 1972), majority language (ML). This process of language shift advances from
generation to generation, yet its speed varies from one community to another depending on
linguistic and social factors. Saltarelli and Gonzo (1977) clarify that the first generation is
bilingual with a strong dominance of the mother tongue, the second generation is bilingual
with a dominance of one language or with a balanced situation, the third generation is
bilingual with a dominance of ML, and the forth generation only masters the ML. This study
investigates language maintenance and shift in three generations of Turkish immigrants in
the Netherlands, the third generation of which is currently reaching adulthood. It examines
the possible change that occurs in language history, self-rated language proficiency and
language use (i.e. Turkish and Dutch) across the generations and the tension between
language maintenance and shift in this community. It should be noted that the concept of
“language tension” used in this study refers to the potential strain experienced between a
socioemotional need to maintain the HL (Canagarajah 2008; Oh and Fuligni 2010) and an
instrumental pressure for full transition to the ML (Portes and Hao 2002). “Language
tension” does not view of bilingualism negatively but rather as a challenge. While the
advantages of bilingualism are often greater than the disadvantages (see Grosjean 1994),
challenges that bilingual families face in an immigrant context (i.e. tension, conflict,
ambiguities) are undoubtedly worth discussing (Baker 2000; Goulbourne et al. 2010).

1.1 The process of language maintenance and shift
The outcome of a language contact situation (i.e. proficiency, language dominance, language
maintenance and shift) is determined by the interplay of both linguistic and social factors
(Gal 1979; Thomason 2008). However, linguists and sociologists often implement different
criteria when investigating these outcomes. Some linguists, for instance, argue the outcome

1
The term ‘heritage language’ in this study is used synonymously with ‘immigrant minority
language’.

2
is determined by the age when language learning begins, while sociologists contend it is
predicted by motivational factors and exposure to the languages in societal contexts
(Stevens 1999). This study considers the significance of age of acquisition of languages (AoA)
and age of arrival in the host country (AoAr) as it is reflected in different generations, but it
emphasizes a number of other social and cultural factors recognized as significant in cases of
language maintenance and shift (Gal 1979).
AoA and AoAr affect how we delineate between early (i.e. child) or late (i.e. adult)
bilinguals. Numerous studies suggest that early bilinguals have certain advantages over late
ones in language learning (DeKeyser et al. 2010). Early bilinguals may be able to construct
two grammatical systems, becoming proficient in both languages (Romaine 1995).
Nevertheless, the degree of proficiency and the pattern of dominance may change over a
lifetime due to changes in language use (Bybee 2010). For instance, if bilinguals begin using
one language more than the other, the less used language, in most cases the HL, will
inevitably become weaker (MacLeod and Stoel-Gammom 2010). Therefore, other factors
such as the frequency with which a language is used, educational attainment in the ML,
social and motivational factors, and family members’ proficiency are vital to consideration in
studies of language maintenance and shift in an immigrant context (Valdés et al. 2006; De
Klerk 2000).
The fate of language maintenance or shift is determined in part by language choice
patterns (i.e. how much and in which domains the HL and the ML are used) (Labov 1972).
Several studies on language choice in social settings have adopted the notion of “social
network” (Milroy 1980; Li 1994; Lanza and Svendsen 2007; Hlavac 2013) to capture domains
such as “family”, “friends”, “work”, “commerce and shops” and “media”. In examinations of
immigrant languages, many studies (Hulsen et al. 2002; Stoessel 2002) conclude that having
social networks within the host community can stimulate language shift. However, due to
social and cultural differences between the ethnic and host communities, immigrants may
fail to construct social networks that include members of the host group, and preserve their
social contact and bond within their ethnic group, in which the HL is often spoken. By
reducing the distance between immigrants and their home country (e.g. through media and
internet), globalization can also support language maintenance efforts of immigrant
communities (Grenoble and Whaley 2006). In such cases, HL maintenance can be sustained
over an extended period of time.

3
 Closely related to language choice and practices, the power relationship between
heritage and majority languages affects the function and dominance of these languages over
the course of heritage speakers’ lives. Thus, even bilinguals who are equally proficient in
both languages can encounter sociolinguistic or emotional pressures that may cause one of
the languages to become more dominant (Montrul 2013). This is particularly evident among
immigrant populations whose language use, proficiency and the degree of linguistic
maintenance or shift are influenced by sociopolitical conditions in the host country and
sociocultural issues, namely their attitudes and their general value system (underpinned by
their identity, culture, religion and so forth) (Edwards 1984). For instance, immigrant families
may be more likely to maintain their HL if they perceive it as an important aspect of their
ethnic identity (Kang and Kim 2012). In such cases, language shift can lead to
intergenerational tension within families (see Portes and Rumbaut 2001), perhaps mostly
when it is associated with identity shift. Thus, the pressure for immigrants to join
mainstream society and the need to resolve intergenerational conflict may lead these
families to forego language maintenance (Canagarajah 2008). It is also important to note
that different immigrant communities may face different language maintenance experiences
due to variations in their general value systems.
In social networks, bilinguals’ gender is also influential, because males and females
often participate in divergent social environments, which impacts immigrants’ levels of
language proficiency. Investigating the complex interrelationship between gender and
language learning, a series of early studies in the field of SLA indicate that women have
superior language abilities compared to men (e.g. Ellis 1994). As opposed to this approach,
other scholars have increasingly adopted a framework that focuses on the role of context,
power relations, ethnic, social, and cultural diversity of gendered identities and linguistic
practices (Piller and Pavlenko 2009; Sharma 2011).
Finally, in addition to social domains, language choice in cognitive activities such as
“emotional expressions”, “mental calculations”, and “inner speech” can tell us much about
language competence and dominance (Dewaele 2004). The link between language
dominance, language preference and emotional/mental representation is discussed in
various studies (Pavlenko 2014). The shift in dominance from the HL is linked to a similar
decline in the use of the HL for the expression of emotions, inner speech and mental
calculations (Dewaele 2004).

4
1.2 The Turkish immigrant context in the Netherlands
As part of general labour migration to Western Europe, the 1960s and early 1970s witnessed
a movement of Turkish immigrants to the Netherlands. Although intended to be short-term,
this migration was followed by family reunifications and migration marriages, substantially
increasing the Turkish population in the Netherlands, which today is among the highest in
the EU countries (Milewski and Hamel 2010). Different components of language
maintenance and shift (e.g. language proficiency, language choice, attitudes towards
languages and cultures) have been investigated in the first and second generations of the
Turkish community in the Netherlands through questionnaires on language skills, language
use and choice patterns (cf. Vedder and Virta 2005; Extra and Yağmur 2010). By means of
conversational data (Doğruöz & Backus, 2009), judgement tasks and controlled experiments
(Doğruöz and Gries 2012; Sevinç 2014), several studies have illustrated that the Turkish
spoken in the Netherlands differs from the Turkish spoken in Turkey both lexically (code-
switching, loan translations) and structurally (Dutch interference). These studies agree that a
gradual language shift seems to be occurring in this community. Furthermore, findings based
on a large-scale survey home language survey Extra and Yağmur (2004) have shown that
even among the second and third generation, Turkish is the most vital immigrant language in
the European context. They have also revealed that Turkish immigrants in the Netherlands
view their HL as a core marker of ethnic identity, and are motivated to maintain it over
generations.
As noted by Extra and Yağmur (2010), first-generation immigrants are Turkish-born
individuals with ties mostly to other members of the Turkish community in the Netherlands,
and they ascribe a high value to maintaining the Turkish language. They frequently visit
Turkey and watch Turkish TV channels. Within this population, the HL tends to be the
dominant language and there is a strong commitment to Turkish identity and culture (Backus
2013). “On the whole, they have functional fluency in Dutch and the number of high
achievers is relatively small” (Yılmaz and Schmid 2015: 124).
For second-generation bilinguals, Dutch was often the weaker language during their
early years in school due to limited use of Dutch at home (Leseman 2000). In this context,
most Turkish-Dutch children started to learn Dutch at school age (Blom 2010). According to
Driessen et al. (2002), even after two years in school, when these children have developed
their Dutch skills, they still lag behind their Dutch classmates. However, after several years of

5
systematic exposure at school, as Extra and Yağmur (2004) suggest, Dutch becomes their
dominant language. In this respect, compared to first-generation immigrants, second-
generation bilinguals have much higher Dutch proficiency, much higher ratings for Dutch
vitality and they also choose to speak Dutch in interaction with others more often (Yağmur
2009).
To a remarkable extent, second-generation bilinguals still choose spouses from Turkey
(Yağmur 2009) and their children are currently forming the third generation of this
community. In the current study, the first generation thus includes two types of immigrants:
those who migrated to the Netherlands through labour migration in the 1960s and early
1970s, and those who migrated through marriage migration after marrying a second-
generation Turkish spouse. The term “second generation” refers to Turkish people who were
born in the Netherlands or arrived before the age of five and started school there2. The term
“third generation” refers to those bilinguals who have one second-generation Turkish-Dutch
bilingual parent and one Turkish-born parent who came to the Netherlands through
marriage migration.

2 The current study
No study to date has compared three generations of this community. Sevinç (2014), a
previous pilot study (n=14), has suggested that with one second-generation parent, the
Turkish language dominance at home decreases substantially in the daily life of third-
generation bilinguals. A larger sample is thus required to consolidate those findings and
other factors influencing language maintenance and shift of the community need
examination. Methodologically, there is also a shortage of research that adopts an
interdisciplinary approach that could give voice to Turkish immigrants’ experiences in the
Netherlands. Therefore, by pursuing an intergenerational comparison and combining

2
According to Statistics Netherlands (2000), which qualified second-generation immigrants based on
the demographic (birthplace) criterion, the second generation was still young (more than 80% is
under 20), and the third generation of Turkish origin was therefore rather small in number (Alders &
Keij, 2001). Unlike Statistics Netherlands, the current study follows sociolinguistic rather than
demographic terminology: the immigrant parents are the first generation, the children second, and
the grandchildren are the third (Silva-Corvalán, 1994). Instead of birthplace, it employs a sociological
criterion to classify the second generation, i.e. whether or not children started their school careers in
the Netherlands. Therefore, the meanings of the terms second and third generation as they are used
in this and other work in contact linguistics are not always identical to how they are defined in official
demographic statistics.

6
questionnaire and interview data, this study aims to explore the following research
questions:
First, is there evidence of the possibility of language shift among third-generation
bilinguals as described in the three-generation language shift pattern? And second, is this
shift also visible in their attitudes towards languages, motivations for learning languages,
opinions about language choice, and immigrant experiences?

2.1 Participants
2.1.1 Questionnaire Respondents
Table 1 provides demographic information for questionnaire participants, 116 Turkish-Dutch
bilinguals 3 (76 female, 40 male); 45 were first-generation bilinguals, 30 were second
generation, and 41 were third generation. They were recruited in Amsterdam, Nijmegen and
Rotterdam, Dutch cities with a sizeable Turkish population. Emphasis was placed on reaching
third-generation immigrants from different socioeconomic backgrounds. Using a snowball
sampling technique, the potential respondents were also requested to ask their family
members to participate in the study.

Table 1: Questionnaire respondents’ demographic information

st nd rd
1 Gen. 2 Gen. 3 Gen.
(n=45) (n=30) (n=41)
Country of Birth NL TR NL TR NL TR
- 100% 64% 36% 100% -
Age (Years) Mean SD (Range) Mean SD (Range) Mean SD (Range)
50 11.7 (32-85) 33 7.3 (20-42) 15 2.8 (11-21)
Gender
Female (n=76) 32 20 24
Male (n=40) 13 10 17
Educational background
Elementary school in Turkey 56%
High school in Turkey 27%
BA in Turkey 17%
High school in the Netherlands 53% 9%
BA in the Netherlands 20%
(Student) MA 10%
(Student) Elementary school 28%
(Student) High school 54%
(Student) University 17% 9%

3
Questionnaires were completed by 131 participants. Ten individuals were excluded since they
acquired Kurdish or Arabic before Turkish and Dutch, and they self-identified as non-Turkish and
non-Dutch. The other five participants were excluded since they were never exposed to Turkish
because they had a Dutch father (and a Turkish mother) and they strongly objected to being
identified as a member of the Turkish community. For the purpose of the study, only those who self-
identified as a member of the Turkish community were included.

7
 Participants ranged in age from 11 to 85. Of the first-generation bilinguals completing
the questionnaire, the majority (56%) completed elementary school and 27% completed
high school in Turkey while 17% graduated from university in Turkey. A greater proportion of
second-generation bilinguals was born in the Netherlands (64%), the rest (36%) came to the
Netherlands before they reached school age. They held either a high school diploma (53%)
or a university degree (20%) from a Dutch institution, while others (27%) were currently
enrolled in a higher education program. All third-generation bilinguals had one second-
generation parent: 52% had one parent born in the Netherlands while the rest had one
parent born in Turkey and came to the Netherlands before school age. Among third-
generation bilinguals, 19% were high school graduates and the majority of them were
currently students: elementary school (28%), high school (54%) and university (9%).

2.1.2 Interviewees
Interviews were conducted with 30 individuals (21 female, 9 male) who had completed the
questionnaires before; 6 were first-generation bilinguals, 8 were second generation, and 16
were third generation (Table 2). Third-generation bilinguals would play the primary role in
the process of language shift. Therefore, the greater sample of third-generation individuals
and their family members was intended to allow for a closer examination of possible
changes at play in the youngest generation.

Table 2: Interviewees’ demographic information

st nd rd
1 Gen. 2 Gen. 3 Gen.
(n=6) (n=8) (n=16)
Country of Birth NL TR NL TR NL TR
- 100% 100% - 100% -
Mean SD (Range) Mean SD (Range) Mean SD (Range)
Age (Years) 41 4.0 (33-43) 26 1.6 (24-28) 14 2.2 (12-19)
Gender
Female (n=21) 6 8 7
Male (n=9) - - 9

All second-generation interviewees were born in the Netherlands. All third-generation
participants had one first-generation parent who came to the Netherlands through marriage
migration, and one second-generation parent who were born in the Netherlands (n=7) or
came to the Netherlands before school age (n=9). Bilinguals from 11 different families were
interviewed; six of these families consisted of siblings and mother, while five of them

8
included only siblings. Although the intention had been to interview both parents in each
family, no fathers wished to participate in the interviews. Interviewees ranged in age from
12 to 43. First-generation interviewees completed their education in Turkey (two were
university graduates, one held high school diploma and three completed elementary school).
Third-generation bilinguals were enrolled either in elementary school or high school in the
Netherlands, while second-generation bilinguals varied in their educational levels from high
school to higher education in the Netherlands.

2.2 Materials
The questionnaire used in this study was designed according to guidelines set by Schleef
(2013). It consisted of four parts: “demographic information”, “language background and
language use and preferences”, “language anxiety”, and “attitudes and experiences”. This
article discusses only the results of the first and second part (see appendix A for the English
version), which comprised questions on bilinguals’ language acquisition history (i.e. when
and where they started to learn languages), language dominance, ongoing language use and
language preferences in six domains (family, school, work, friends, media use and leisure
time, and cognitive activities), as well as bilinguals’ self-reports of proficiency in four skills in
both languages. The first two parts of the questionnaire were adapted from two sources: the
bilingualism and emotions questionnaire (Dewaele and Pavlenko 2001-2003) for the scales of
language proficiency and dominance and language use in cognitive activities, and the
language use and maintenance (GB) questionnaire (Jamai 2008) for the scales of language
use and preferences.
Regarding language dominance, participants were asked which language(s) they
consider their dominant language, which they are most fluent in and most comfortable
speaking. They could choose one of three options: Turkish, Dutch, or both (i.e. Turkish and
Dutch). Participants were asked to choose a value from a 5-point scale for the amount of
language use in six domains (1=Never, 2=Rarely, 3=Sometimes, 4=Frequently, 5=All the time)
and self-reported language proficiency (1=None, 2=Poor, 3=Fair, 4=Good, 5=Excellent). Table
3 contains score reliability information for the scales of language proficiency and language
use in which Cronbach’s Alpha coefficients vary from .69 (acceptable) to .96 (very
satisfactory).

9
Table 3: Reliability of the language proficiency and language use scales
The scale Number of items Cronbach’s Alpha
Dutch Turkish
Language proficiency 4 .96 .95
Language use in domains
Home/family domain 6 .71 .78
Outside 3 .69 .72
Media and leisure time 9 .88 .88
Cognitive activities 3 .91 .90

Semi-structured, open-ended interviews were also conducted (See appendix B for the
English translation). The interviews were designed to delve deeper into bilinguals’ language
history, linguistic dominance, and language practices and choices, while also providing a
means to explore the story behind their experiences, learning languages motivations and
attitudes towards bilingualism.

2.3 Procedure
The questionnaire was administered in the language of the participants’ choice: Turkish or
Dutch. Clear instructions on how to complete the questionnaire were given both in Turkish
and in Dutch. At least one researcher was present at the time that participants completed
the questionnaire, and was therefore available to help if they had any questions.
Respondents finished the questionnaire independently in approximately 25 minutes, 89% of
first-generation, 57% of second-generation and 10% of third-generation participants chose to
fill in the questionnaire in Turkish.
Interviews were held one month after the questionnaire data had been collected. Each
interview lasted between forty minutes and one hour. Interviewees were given clear
information about the content of the interview questions beforehand. They were informed
that their language performance would not be evaluated or scored and they could mix their
languages freely during the interviews if they had difficulties in explaining themselves
without switching to Turkish or Dutch. There were two interviewers, one of Turkish origin,
the other Dutch. In order to prevent a possible interlocutor effect, the subjects were asked to
choose their interviewer themselves by considering the criterion that they could share their
personal experiences more comfortably and honestly. The rational here was that giving
interviewees freedom of selecting their interviewer might help create a stress-free
atmosphere, and thus provide more reliable information. Twelve chose the Dutch researcher

10
and 18 chose the Turkish researcher. The researchers had participants’ previous responses to
the questionnaire, and could therefore inquire into the remarks participants made earlier.

2.4 Analysis
For each item on the questionnaire, responses were coded as continuous (e.g. self-rated
proficiency levels and frequency of language use) or categorical (e.g. generation, dominant
language, place of language acquisition). Descriptive statistics were used to summarize the
results in terms of frequencies and percentages for categorical variables as well as means
and standard deviations for numerical variables. For the purpose of this study, focus was
placed on cross-generational comparisons pertaining to the three main categories: language
proficiency levels, language dominance, and daily language use. In order to reduce the large
number of language background variables, generation effect were reported only on self–
reported overall daily language use, and averaged scores of the four skills of language
proficiency in Turkish and Dutch, rather than on each subcategory (e.g. language proficiency
in four skills or language use in six different domains). The assumption of normality was
checked, a series of one-sample Kolmogorov-Smirnov tests revealed non-normal
distributions for self-rated overall language proficiency levels of Turkish and Dutch and self-
reported overall daily language use (both significant at p < .005). Consequently, in order to
capture the potential group effects (generation) on these variables, Kruskal-Wallis one-way
analysis of variance was used as nonparametric equivalents to one-way ANOVA (see Field
2013). After Kruskal-Wallis tests, follow-up Mann-Whitney U tests were applied to examine
unique pairs. In order to discover relationships between two categorical variables
(generation, dominant language, gender and educational background), Pearson Chi-square
test and Cramer’s V statistics–an indicator of effect size–were used.
The interview data was transcribed and attention was given to participants’
experiences of language maintenance and shift across generations. Procedures for “open
coding”, i.e. the process of breaking down, examining, comparing, conceptualizing and
categorizing data (Strauss and Corbin, 1990), were applied to provide structure and to give
overviews of the interview texts. Instances of the coded transcripts and themes were
reviewed and compared for similarities and differences by two researchers, more focused
coding was then undertaken. Following the merging data approach (Creswell and Plano Clark
2011), excerpts from the coded transcripts were selected to further investigate certain

11
findings from the questionnaires, including additional details that could not be obtained
through questionnaires alone.
In this study, the quantitative statistical results are followed by the qualitative quotes
that support or refute the conclusions drawn from the quantitative data. Findings based on
the questionnaires are presented in four main categories: (1) Language history (i.e. AoA, and
place of language acquisition), (2) language proficiency, (3) language dominance (language
dominance and gender, language dominance and education), and (4) language use and
choice (daily use of languages, and use of languages in six domains). Finally, findings based
on interviews are discussed in three categories: (1) Prior exposure to languages, (2) use of
languages and language dominance, and (3) attitudes and motivation.

3 Results
3.1 Findings based on questionnaires
3.1.1 Language history across generations
Table 4 presents the age and place of language acquisition for participants. Except for four
third-generation bilinguals, all respondents had acquired Turkish from birth, reporting they
were most often exposed to Turkish in a familial context at home. Without exception, all
first-generation immigrants learned it in Turkey, whereas most second-generation and all
third-generation bilinguals learned it at home in the Netherlands. Twenty-five second-
generation bilinguals were exposed to Turkish before family reunification, and one attended
Turkish language courses in the Netherlands.

Table 4: Language history of Turkish-Dutch bilinguals across generations

Dutch Turkish
st nd rd st nd rd
1 Gen. 2 Gen. 3 Gen. 1 Gen. 2 Gen. 3 Gen.
Age of acquisition M SD M SD M SD M SD M SD M SD
22.0 7.2 3.6 2.9 1.1 1.7 0.0 0.0 0.0 0.0 1.0 2.2

Place of acquisition
Home in Turkey 100% 25%
Home in the Netherlands 16% 59% 72% 100%
Kindergarten 22% 24%
School 62% 17%
Language course 49% 3%
Work 11%
Self-study 40%

Age and place of Dutch acquisition differ considerably across the three generations. All

12
of the first-generation immigrants acquired Dutch after puberty, that is, as late bilinguals.
Almost half of them attended a Dutch language course, while the rest acquired it through
self-study or at work. The majority of second-generation bilinguals were exposed to it as
school-aged children after they acquired Turkish. Twenty-two percent acquired Dutch as
kindergarten children, while 16% began acquiring it from birth in their home environment.
The majority of third-generation children reported being exposed to Dutch at home from
birth. Of the rest, 24% percent were exposed to Dutch at pre-school age and 17% as school-
aged children.

3.1.2 Language proficiency across generations
A Kruskal-Wallis test showed a highly significant generational effect for self-rated overall
Dutch proficiency (χ2 = 70.62, df = 2, p < .001), and Turkish proficiency (χ2 = 57.87, df = 2, p
< .001). Third-generation bilinguals reported higher proficiency in Dutch than first- and
second-generation bilinguals. Conversely, third-generation bilinguals had lower self-rated
proficiency in Turkish than first- and second-generation bilinguals. The means and standard
deviations on bilinguals’ self-reported overall proficiency in Dutch and Turkish are presented
in Table 5, which indicate a gradual decline of language proficiency levels in Turkish and a
contrasting increase in Dutch from generation to generation. This in turn points to a possible
ongoing language shift.

Table 5: Self–reported language proficiency across generations

Dutch Turkish
st nd rd st nd rd
1 Gen. 2 Gen. 3 Gen. 1 Gen. 2 Gen. 3 Gen.
Language proficiency M SD M SD M SD M SD M SD M SD
3.3 0.8 4.6 0.4 4.9 0.4 4.7 0.4 4.2 0.5 3.1 0.9
Note. Range of language proficiency: (1) None, (2) Poor, (3) Fair, (4) Good, (5) Excellent

Figure 1 summarizes the self-reported proficiency in both languages for four skills
across the three bilingual groups. First-generation participants reported the highest
proficiency in Turkish and the lowest proficiency in Dutch in the four skills of understanding,
reading, speaking and writing. Importantly, both second- and third-generation bilinguals
rated their Turkish proficiency levels lower than Dutch. They reported the lowest Turkish
proficiency in the skills of reading and writing, which can be caused by lack of systematic
educational experiences in the HL.

13

Figure 1: Self-reported Dutch and Turkish proficiency in four skills

3.1.3 Language dominance across generations
A Pearson Chi-square test showed a main effect of generation on bilinguals’ self-rated
language dominance (χ 2 ( 2 ) = 94.3, p < .001), with third-generation bilinguals being
significantly more dominant in Dutch than first- and second-generation bilinguals. For this
chi-square test, Cramer’s V shows the effect size as .63 (with df=4), which is categorized as a
large effect (Cohen, 1992). Table 6 summarizes bilinguals’ self-reports on the language(s)
they considered to be their dominant language(s), which they were most fluent in, and which
they spoke most comfortably.

Table 6: Language dominance in three generations of Turkish community in the Netherlands

st nd rd
Dominant language 1 Gen. 2 Gen. 3 Gen.
Both 13% 43% 7%
Dutch 0% 37% 88%
Turkish 87% 20% 5%

These findings reveal a sharp contrast among generations in terms of language

dominance. Turkish as a dominant language among first-generation immigrants yields to
Dutch among third-generation bilinguals. In comparison to second-generation bilinguals, the
number of bilinguals who reported being dominant in two languages also decreases by the
third generation. Hence Dutch, the ML, seems to have become the dominant language of the
youngest generation.

3.1.3.1 Language Dominance and Gender
A Pearson Chi-square test found an effect of gender on dominant language (χ 2(2) = 6.82, p

14
< .05). The males to a large extent chose Dutch as their dominant language (Table 7). For this
chi-square test, Cramer’s V shows the effect size as .24 (with df = 2), which is categorized as a
medium effect size (Cohen, 1992). Considering that the gender bias is great in the participant
sample, this interaction must therefore be treated with caution.

Table 7: Participant's gender and dominant language

Gender
Dominant language Female Male Total
Both 18% 20% 20%
Dutch 32% 55% 40%
Turkish 50% 25% 40%

3.1.3.2 Language dominance and educational background
A Pearson Chi-square test also found an effect of educational background on dominant
language (χ 2(16) = 94.32, p < .001). For this chi-square test, Cramer’s V shows the effect size
as .64 (with df = 16), which is categorized as a large effect size (Cohen, 1992). Table 8
illustrates that the participants who went to school in the Netherlands mostly selected Dutch
as their dominant language while the ones who went to school in Turkey mostly chose only
Turkish as their dominant language.

Table 8: Participants’ educational background and dominant language

Educational Background
Dominant language Elementary High school BA High school BA (HBO) Student Student Student Student
in TR in TR in TR (HAVO) in NL in NL Elementary HAVO HBO MA
Both 12% 9% 5% 46% 5% 0% 5% 18% 0%
Dutch 0% 0% 0% 17% 6% 24% 40% 9% 4%
Turkish 48% 21% 15% 4% 4% 0% 4% 2% 2%
Note. In the Netherlands, the HAVO (hoger algemeen voortgezet onderwijs) literally, "higher general continued education" has five grades
and is attended from age twelve to seventeen. A HAVO diploma provides access to HBO, the level of (polytechnic) tertiary education.

3.1.4 Language use and choice

The questionnaire elicited information on bilinguals’ daily use of languages, and use of
languages in six selected settings (family, friends, school, work, media and leisure time, and
cognitive activities).
With regard to overall daily language use across generations, Kruskal-Wallis test
revealed a main effect of group in self-reported daily language use: Dutch (χ2 = 39.34, df = 2,
p < .001), Turkish (χ2 = 70.96, df = 2, p < .001) and mixing Dutch and Turkish (χ2 = 15.24, df =
2, p < .000). Third-generation bilinguals reported that they used Dutch more frequently than
first- and second-generation bilinguals. Third-generation bilinguals also reported using

15
Turkish less frequently than first- and second-generation bilinguals. Both third- and second-
generation bilinguals mixed Dutch and Turkish more frequently than first-generation
immigrants. Note that follow-up Mann-Whitney U test showed no significant differences
between first- and second-generation bilinguals on daily Turkish use (U = 54, p > .05), and
between second- and third generation bilinguals on daily amount of mixing Dutch and
Turkish (U = 53, p > .05).
Table 9 summarizes bilinguals’ self-reports on their daily language use. Third-
generation bilinguals’ seem to use Dutch all the time. The majority of this generation
reported using Turkish rarely or sometimes, while mixing Dutch and Turkish sometimes or
frequently. Daily use of Dutch and mixing the languages was the least frequent among first-
generation immigrants and use of Turkish the most frequent. As for second-generation
bilinguals, although use of Dutch was the most frequent, Turkish and mixing languages also
seemed to be a major part of their daily language use. These findings suggest a decline in the
daily use of Turkish and an increase in Dutch use among the third-generation group, which
can therefore be evidence for a possible language shift.

Table 9: Self–reported daily language use across generations

Dutch Turkish Mixing
st nd rd st nd rd st nd rd
1 Gen. 2 Gen. 3 Gen. 1 Gen. 2 Gen. 3 Gen. 1 Gen. 2 Gen. 3 Gen.
Daily M SD M SD M SD M SD M SD M SD M SD M SD M SD
Language use 3.3 1.2 4.3 0.9 4.8 0.4 4.6 0.6 4.2 0.8 2.6 0.7 2.8 1.0 3.7 1.0 3.5 1.0
Note. Range of language use per day: (1) Never, (2) Rarely, (3) Sometimes, (4) Frequently, (5) All the time

Results on language use will now be presented in six domains (i.e. home, school, work,
friends, media and leisure time activities, and cognitive activities). With regard to language
use in the home/family domain, Table 10 presents language use at home. Third-generation
bilinguals seem to use a great deal of Dutch when speaking with their siblings, more Dutch
than Turkish with their mothers, and more Turkish than Dutch with their fathers and
grandparents. High standard deviations within the groups indicate that the amount of
Turkish and Dutch used at home can vary depending on the composition, language
background, and attitudes of each family. Nevertheless, it is evident that in comparison to
first- and second-generation bilinguals, there is a marked increase in the amount of Dutch
that third-generation children use at home.

16
Table 10: Self–reported language use in the home/family domain

Dutch Turkish Mixing
st nd rd st nd rd st nd rd
1 Gen. 2 Gen. 3 Gen. 1 Gen. 2 Gen. 3 Gen. 1 Gen. 2 Gen. 3 Gen.
M SD M SD M SD M SD M SD M SD M SD M SD M SD
1. Home/family 2.3 1.3 3.2 1.2 4.2 1.0 4.6 0.7 4.2 0.8 3.4 1.0 2.6 1.1 3.2 1.1 3.5 1.1
With mother 1.7 0.9 3.5 1.4 4.9 1.0 4.8 0.6 3.2 1.3 1.9 1.2 3.1 1.3
With father 1.6 1.0 2.9 1.5 4.9 0.4 4.8 0.5 3.6 1.4 1.9 1.3 2.7 1.3
With siblings 1.4 0.9 3.2 1.2 4.7 0.8 4.7 0.5 4.0 0.8 2.3 1.2 1.4 0.9 3.7 1.0 2.8 1.4
With children 2.7 1.5 3.8 1.2 4.2 1.0 3.9 1.0 2.6 1.1 3.2 1.2
With partner 2.1 1.6 2.9 1.4 4.6 0.7 4.4 1.2 4.2 0.9 2.7 1.5 1.7 1.1 2.8 1.4 3.1 1.7
With grandparents 1.9 1.5 4.9 0.5 3.9 1.4 1.3 1.1 1.4 0.8
Note. Range of language use: (1) Never, (2) Rarely, (3) Sometimes, (4) Frequently, (5) All the time

Another striking difference between generations is the shift in language use with
partners. Unlike first- and second-generation participants who speak Turkish very often with
their spouse, third-generation bilinguals reported using much more Dutch than Turkish with
their girlfriend/boyfriend. Note that 40% of second-generation immigrants married a Turkish-
born spouse, while 17% married a Dutch-born person with Turkish ethnic background and
43% were either single or divorced. On the other hand, 36% of third-generation bilinguals
had Dutch-born partners (girlfriends/boyfriends) while the rest were single.
With regard to language use outside across the domains of friends, school and work,
Table 11 summarizes information on bilinguals’ self-reported language use across the
domains of friends, school and work. The data suggest an increased use of Dutch among
third-generation bilinguals. While first-generation immigrants speak Turkish frequently or all
the time with their friends, Turkish seems to be replaced by Dutch in this setting for third-
generation bilinguals. Second-generation participants reported a frequent use of Turkish and
frequent mixing of Turkish and Dutch with friends, whereas third-generation bilinguals
reported using Turkish never or rarely and mixing the languages rarely or sometimes.

Table 11: Self–reported language use across the domains of friends, school and work

Dutch Turkish Mixing
st nd rd st nd rd st nd rd
1 Gen. 2 Gen. 3 Gen. 1 Gen. 2 Gen. 3 Gen. 1 Gen. 2 Gen. 3 Gen.
M SD M SD M SD M SD M SD M SD M SD M SD M SD
2. Friends 2.7 1.3 4.2 0.7 4.7 0.6 4.3 0.9 4.0 8.8 2.4 1.1 2.4 1.1 3.6 1.0 2.8 1.4
3. School 4.8 0.5 4.9 0.2 2.1 1.2 1.4 0.8 2.2 1.3 1.7 1.1
4. Work 4.5 1.0 4.8 0.5 4.8 0.4 2.3 1.7 2.4 1.4 2.3 1.4 2.1 1.4 2.0 1.1 2.6 1.4
Note. Range of language use: (1) Never, (2) Rarely, (3) Sometimes, (4) Frequently, (5) All the time

Among the other settings, the school domain (for second- and third-generation
children) and work domain (for all generations) displayed the highest amount of Dutch and

17
lowest amount of Turkish use. Not surprisingly, second- and third-generation bilinguals used
Dutch most often in the context of school and bilinguals of all three generations in the
context of work.
With regard to language use across domains of media and leisure time, and cognitive
activities, an overview of bilinguals’ self-reported language use and preferences based in
two other domains, media/leisure time, and cognitive activities, are presented in Table 12.

Table 12: Self–reported language use and preferences across the domains of media and
cognitive activities

Dutch Turkish
st nd rd st nd rd
1 Gen. 2 Gen. 3 Gen. 1 Gen. 2 Gen. 3 Gen.
M SD M SD M SD M SD M SD M SD
5. Media/leisure time
Books/novels 1.8 1.0 3.0 1.5 4.5 1.0 3.5 1.6 2.6 1.3 1.4 0.8
Newspapers 2.2 1.2 3.6 1.4 3.8 1.5 3.2 1.5 2.8 1.4 1.2 0.6
Magazines 2.0 1.2 3.0 1.6 4.0 1.3 2.9 1.6 2.3 1.4 1.1 0.5
Radio 2.2 1.4 3.2 1.6 4.0 1.3 3.1 1.8 2.4 1.5 1.9 1.3
Television 2.4 1.3 3.3 1.4 4.0 1.2 4.5 0.8 4.1 1.0 3.0 1.5
Films/TV series 2.2 1.3 3.0 1.5 3.8 1.4 4.4 0.9 3.9 1.3 3.0 1.6
Music 1.9 1.3 2.5 1.5 2.9 1.7 4.2 1.2 4.0 1.2 2.8 1.5
Computer 2.8 1.7 4.6 0.8 4.5 1.0 3.4 1.7 2.7 1.6 1.8 1.1
Social media 2.4 1.3 4.0 1.1 4.6 0.6 3.6 1.5 3.5 1.5 2.2 1.2
6. Cognitive activities
Emotions 2.2 1.3 3.6 1.1 4.5 0.7 4.6 0.7 4.2 0.8 2.4 1.2
Arithmetic 1.9 1.3 4.2 0.8 4.9 0.3 4.7 0.6 3.5 1.1 1.3 0.7
Inner speech 1.9 1.1 3.2 1.1 4.4 0.9 4.6 0.8 4.2 0.8 2.5 1.3
Note. Range of language use: (1) Never, (2) Rarely, (3) Sometimes, (4) Frequently, (5) All the time

Cross-generational differences in language use while using media (especially TV and
internet) and engaging in leisure time activities provide further evidence of the decline in
Turkish use among third-generation bilinguals. The majority of these young bilinguals
reported they almost never read books/novels, newspapers or magazines in Turkish, never
listened to Turkish radio and only rarely used computer and social media in Turkish. On the
other hand, the frequency of third-generation children reading books, newspapers, and
magazines in Dutch and listening to Dutch radio as well as using computer and social media
in Dutch are relatively higher than their self-rated frequency levels for Turkish. Meanwhile,
listening to Turkish or Dutch music is less frequent among this group, as they mostly listen to
songs in English.
To a great degree, Turkish TV, Turkish films and TV series occupy more leisure time of
first-generation immigrants and, less frequently, second-generation bilinguals, than Dutch TV.
On the other hand, third-generation children watch more Dutch TV than Turkish.

18
 Finally, according to self-assessments of third-generation participants, Dutch is also
their preferred language for emotional communication, mental calculation and inner speech,
Turkish being used very rarely for these cognitive activities. First-generation immigrants
showed the opposite picture. Second-generation bilinguals reported using more Turkish than
Dutch for emotional expressions, as well as for inner speech, yet more Dutch than Turkish for
calculations.
Findings based on bilinguals’ language use and preferences suggest that Dutch has
become the dominant language of many young bilinguals in many or all domains. The
predominant use of Turkish in each domain remains strong among first-generation
immigrants, while it varies among second-generation and diminishes for third-generation
bilinguals.

3.2 Findings based on interviews
The questionnaire data show that three generations of Turkish immigrants differ in their
language history, self-reported proficiency and use patterns. Overall, the interviews support
the questionnaire findings on the gradual decline of Turkish use from generation to
generation. Yet they also provide us with more detailed information about the causes of
language shift, and the variation of immigrant experiences, that exists not only among
different generations but also within a generation and even within a family.
The key findings from the interviews are presented below in three categories: 1) prior
exposure to languages, 2) current exposure to languages and language dominance and 3)
motivation for and attitudes toward learning languages. Special emphasis is placed on
experiences that influence the factors that bilinguals themselves see as contributing to
language maintenance or shift.

3.2.1 Prior exposure to languages
Interviewees raised a number of key issues that help explain the composition of their
repertoires. First of all, for many first-generation immigrants in this community, particularly
for women, age of arrival in the Netherlands does not necessarily mean age of exposure to
the Dutch language. Three of the six first-generation women interviewed (AA, SDD, DG) were
divorced, and they claimed that Dutch acquisition for them had begun post-divorce. DG (33
years old) describes her experiences as follows:

19
 Excerpt (1)
Interviewer: When and how did you arrive in the Netherlands?
DG: Eighteen years ago, when I was 15. It was never my own decision. I was forced to
marry and brought here… Later on, the kids were born, I took care of them, only my ex-
husband’s family came for a visit, we talked Turkish all the time.
Interviewer: When did you learn Dutch then?
DG: Six years after [arrival]. Right after I divorced my husband. Before the divorce, I
didn’t speak even one word in Dutch. Then, I went to a course, I practised it with my
children at home, Dutch neighbors...

One of the first-generation participants (AT, 42 years old) explained how she was
subject to her husband’s authority which included restrictions on her Dutch language
practices. AT had been in the Netherlands for 26 years and had almost no command of
Dutch. When asked why, she explained:
Excerpt (2)
AT: My husband didn’t let me learn it; I couldn’t go to language courses… He’s scared
that I’ll divorce him if I speak Dutch. He doesn’t want me to learn Dutch, because he
just wants me to be in need of him…

AT’s situation illustrates that even after many years in the Netherlands, Turkish
immigrants may have limited Dutch knowledge due to family-related restrictions. It also
suggests that knowing Dutch carries some power that is evident not only in political, social
and economic contexts, but also in the contexts of immigrant families. A good command of
Dutch is equated with power, equality and freedom for some women while it denotes
domination to some men.
As for second-generation bilinguals, all of them grew up in minority neighborhoods
where their parents often socialized with other Turkish families. They were not exposed to
Dutch until they started school. SAK (27 years old) describes her first day of school:
Excerpt (3)
SAK: I’ll never forget the first day of school. I couldn’t even speak one word of Dutch.
The teacher was talking and I didn’t understand one word. I was scared to death of
speaking Dutch. I couldn’t make friends for a long time, I felt like an idiot. At home,
there was nobody who could help me with my homework. For instance, my youngest
sister never experienced these difficulties as I did. I always helped her and spoke Dutch
to her. Thus, her Dutch is better, but she is ashamed of her Turkish...

SAK’s statement also points out the variation within one family; where younger siblings
may be exposed earlier to Dutch and with more intensity.
Unlike second-generation bilinguals, many third-generation children indicated that they
acquired Turkish and Dutch from birth. In addition, they sometimes mentioned “the

20
neighborhood effect,” or as DG described it, “practicing Dutch with Dutch neighbors”. Many
participants emphasized the effects of their neighborhood on their language knowledge. Two
third-generation bilinguals, PK (18 years old) and EK (14 years old) indicated that they were
exposed to Turkish only after moving to a neighborhood with predominantly minority
families, and they started to play with other Turkish immigrant children. Here is PK's
description:
Excerpt (4)
Interviewer: When and how did you learn the languages?
PK: First actually Dutch and later Turkish. But that was really years later.
Interviewer: So at home Dutch was spoken?
PK: Yes, mostly Dutch...We also lived in a neighborhood where only Dutch people lived
so it was impossible to learn Turkish…Later, we moved to a place where there were
more Turkish people…
Interviewer: How old were you when you started to speak Turkish?
PK: 14, 15…

3.2.2 Use of languages and language dominance
When asked what the prevalent language was in their current home, first-generation
immigrants answered “Turkish occasionally mixed with some Dutch words”. Second-
generation bilinguals reported “a mix of Turkish and Dutch,” while third-generation bilinguals
said “Dutch” or “mixing languages” with parents and “mostly Dutch” with siblings. Third-
generation children whose dominant language is Dutch may still be exposed to a
considerable amount of Turkish through contact with parents or other older family members,
yet they feel more comfortable responding in Dutch even when addressed in Turkish. This is
how DG, a first-generation immigrant married to a second-generation man, described her
third-generation daughter’s Turkish:
Excerpt (5)
DG: Her Turkish is terrible. Just now too, I was talking Turkish outside, she didn’t
understand me, she was staring blankly. Even if she understands, she doesn’t respond in
Turkish. Because she finds Dutch easy, and automatically she starts speaking Dutch. In
fact, she knows [Turkish] but when she doesn’t practice…I don’t feel that she knows it in
practice... I tell her “read Turkish books, watch Turkish channels,” she doesn’t do any of
that.

DG’s daughter, 13-year-old IK elaborated on her interaction with her mother (DG)
regarding language use, also noting the tension it created:
Excerpt (6)

21
 IK: My mother gets angry when I don’t understand her. She says “your Turkish is too
bad, read Turkish books, watch Turkish channels”. Then we quarrel. Then I don’t speak
Turkish with her. She asks in Turkish, I reply in Dutch.
Interviewer: When do you speak Turkish then?
IK: With my friend after school.
Interviewer: How do you speak? Only Turkish?
IK: No. We can’t speak only Turkish. Because we don’t know many things {words} in
Turkish. We make mixed {mixing Turkish and Dutch words}.

On the other hand a 14-year-old, third-generation ET illustrated the tension between
her and her grandfather as follows:
Excerpt (7)
ET: My grandfather makes me stressed about my Turkish. He says that I can’t talk
Turkish very well. I should fix it, he says, otherwise I won’t be able to find a husband. He
even says even if I find one [husband], I won’t be able to communicate with my mother-
in-law.

Another conflict regarding language use at home arises when spouses who move from
Turkey use their new home environment to practise their Dutch with their children. This
trend of learning Dutch from/with children seems to reduce the amount of Turkish exposure
at home whilst also causing tensions in the family about Dutch proficiency. According to 13-
year-old third-generation SE, his mother, a second-generation bilingual, insisted her children
not speak Dutch with their father (a first-generation immigrant) because speaking Dutch with
him negatively influenced the children’s Dutch proficiency. This is how he explained the
situation:
Excerpt (8)
Interviewer: Are there any rules at home you have to obey regarding speaking a
language?
SE: Yes, actually I have to speak Turkish to my dad. Because my mother says I also
started to speak poor Dutch because of him, because I copied my dad’s language when
he spoke Dutch, wrong use of language and incorrect syntax.

The other third-generation interviewees said that there were no rules about language
use at home nor were there extra activities to improve their Turkish. For instance, they
neither read Turkish books nor had someone read books to them when they were little.
According to mothers interviewed, the only Turkish practice tool at home was Turkish
satellite TV but their children preferred Dutch TV. These language choices reflect young
bilinguals’ increasing dominance in Dutch.

22
3.2.3 Attitudes and motivation
In the interviews, children and mothers alike voiced very positive attitudes about
multilingualism and speaking correct Dutch and Turkish. As they saw it, knowing Dutch leads
to economic security, the ability to communicate and interact with a wide range of people,
and access to knowledge sources outside of their immediate community. Turkish, on the
other hand, helps connect them with Turkish culture and their relatives, particularly those
back in Turkey.
First-generation immigrants were aware of the immediate advantages of having
children who could translate for older family members helping especially to communicate
with Dutch authorities. All six first-generation bilinguals expressed the importance of
maintaining Turkish by saying “because it is my mother tongue”. Three of the first-generation
immigrants emphasized their strong emotional attachment to Turkey, Turkish culture, and
indeed religion, Islam. AT (42 years old), a first-generation mother with almost no command
of Dutch, made this aspect clear:
Excerpt (9)
AT: For me, Turkish is more important because we are Muslims, it is also my own
mother tongue... Dutch is important as well, as you need to make yourself understood
here [in the Netherlands]. For instance, I can’t even go to the doctor without my
daughters, they act as an interpreter for me… And also, Dutch is important for the next
generation to reach a higher status in the Netherlands. They must know Dutch very well
without losing our own language (Turkish) and culture.

As mentioned by AT, first and second generation bilinguals expressed their expectations
for their children and grandchildren to maintain the Turkish language and culture, at the
same time, to master the Dutch language completely.
Like first- and second-generation bilinguals, the majority of third-generation children
thought that Turkish was important because it was linked to Turkish identity and culture. In
the words of SE:
Excerpt (10)
SE: Turkish is important, and I have to speak it perfectly because it’s in my blood...[It’s]
more like being proud of your country…loving your own country, you know. Because
every person loves his/her own country.

For SE’s older brother, 15-year-old IE, Turkish is less important, but he had to learn it
for the sake of his culture:

23
 Excerpt (11)
IE: Dutch is more important because I live in the Netherlands. I think Turkish is kind of
less important to me. But it still belongs to my culture. So I have to know Turkish.

Only a few second- and third-generation bilinguals identified themselves as neither

Turkish nor Dutch. They indicated that sometimes they felt caught between two cultures.
Nevertheless, they all emphasized that both languages were important. This is how 26-year-
old, second-generation SLD explained it:
Excerpt (12)
Interviewer: You said your Turkish is not so perfect. Do you think you need to speak
Turkish “perfectly”?
SLD: Sure.
Interviewer: Why is that?
SLD: …Even though I was born in the Netherlands, my parents are Turkish. Actually, I
don’t feel like either a Turkish or a Dutch person. It’s so bad that we’re mixed up in
between two countries... However, the languages are important. Dutch is a must, in
order not to let Dutch people overpower you, in order to prove yourself to Dutch
people. Turkish is also important, in order not to be fooled by Turkish shopkeepers when
we do shopping in Turkey. They say a higher price when they find out we live in Europe4.

All participants were well aware of and concerned about the decrease in their use and
knowledge of Turkish. Regarding opinions about language choice, all the third-generation
bilinguals indicated they preferred speaking Dutch because they thought their Turkish was
not very good. KB, a 15-year-old third-generation bilingual, makes this clear:
Excerpt (13)
KB: I also try to speak Turkish sometimes, but it isn’t that great… So I prefer to speak
Dutch. Or just Dutch sentences and then something in Turkish or so.

The majority of the second- and all third-generation bilinguals indicated that the
Turkish they spoke was entirely different from the Turkish spoken in Turkey. Some thought
that their Turkish was bad due to the influence of Dutch, some thought it was due to
excessive exposure to Dutch or due to the frequent intermingling of Turkish and Dutch. SVD,
a 26-year-old second-generation immigrant, assessed the Turkish spoken in the Netherlands
and in Turkey and elaborated on the consequences of Turkish-Dutch language contact (i.e.
influence of Dutch, code-switching):
Excerpt (14)
SVD: For instance, I can’t speak comfortably in Turkey, I don’t know, I mean people
there sound too kind and modern to me… We, the Turks in the Netherlands, talk pell-
4
This excerpt, as many others, suggests a possible internalized discrimination that the Turkish
immigrants living in Europe face in Turkey. This will be further elaborated in a future study.

24
 mell [messily]. I don’t know if it is because of the “gh” sound of Dutch; I mean we speak
rudely… And also, we’re not used to speaking only Turkish; one of every two words we
speak is Dutch, this certainly has an effect [on it] too.

Despite the value they ascribed to Turkish language and culture, the majority (22) of
second- and third-generation bilinguals stated that Dutch occupied a significant place in their
lives. They emphasized that Dutch was more important than Turkish since they were born
and lived in the Netherlands. All third-generation bilinguals were aware that succeeding in
school and a career required speaking “perfect Dutch”.
25 of the participants in all generations stressed that they experienced pressure from
the Dutch community to speak Dutch (e.g. at school, at work). Furthermore, they were all
concerned that speaking Turkish was not tolerated particularly at schools. ST (11 years old)
elaborated on the consequences of speaking Turkish at school:
Excerpt (15)
ST: You can’t speak Turkish at school, there are punishments, you are writing.
Interviewer: What are you writing?
ST: You write in Dutch… Writing homework punishment.

Six second-generation bilinguals shared their concerns and experiences of

discrimination. All of them agreed that they had to speak Dutch as well as, or better than a
Dutch person to avoid being discriminated against or alienated at school or at work. For
many, the Dutch language was a key factor in proving themselves to Dutch people and
leading a satisfying life without being despised due to their different cultural background.
SLD’s recounting of discrimination she experienced due to language limitation summarizes
the prejudice faced by second-generation children at school:
Excerpt (16)
SLD: …I experienced that when someone goes to school, if she doesn’t speak Dutch very
well, they point their fingers at her, [it’s] more like they expel her from everything. For
instance, they mostly talk to the ones who speak Dutch better. My teacher, for instance,
didn’t like the students who come from outside [of the Netherlands]… He always put us
down saying, “they are Turks, Moroccans, how do they know it…” So, I don’t want my
child experiencing these things as well. That’s why for me Dutch is more important.

Based on their experiences, second-generation bilinguals challenge the trend that “the
kids will learn Dutch at school anyway”. This situation may influence the amount of Dutch
and Turkish used at home and also the age of acquisition of these languages.

25
4 Discussion and concluding remarks
Given that the current findings are based on self-reports and there is gender bias in the
current sample, the interpretations regarding language shift and gender differences
presented here should be treated with caution. Despite these limitations, this study sheds
crucial light on a possible language shift occurring among third-generation Turkish
immigrants in the Netherlands.
Questionnaire and interview data demonstrate that a shift from Turkish to Dutch
might be occurring from generation to generation. This was evident in self-rated language
proficiency, language dominance and for language choice in six domains: home, school,
work, friends, media and leisure time activities, and cognitive activities. Compared to first-
generation immigrants, the use of Turkish varies more among second-generation, but
declines considerably among third-generation bilinguals.
Overall, language history and current language practices of third-generation bilinguals
certainly differ from first- and second-generation bilinguals. As suggested in Sevinç (2014),
Dutch has gained more ground among these young bilinguals, and having only one second-
generation parent at home has increased their exposure to Dutch. Following the pattern of
language shift described by Saltarelli and Gonzo (1977), the amount and dominance of Dutch
in third-generation bilinguals’ daily lives have increased, and acquisition of Dutch has begun
occurring before school age or from birth. Globalization, which supposedly promotes
language maintenance, does not seem to stop language shift. The predominant use of Dutch-
language media and internet among third-generation bilinguals’ indicates a perceived
tension and need to shift to the ML may prevail against language maintenance efforts.
The findings from the interview data show that AoAr may not be a definite factor in
determining Dutch proficiency levels of first-generation Turkish immigrants, particularly for
women. As in recent studies on gender and multilingualism (Piller and Pavlenko 2009;
Sharma 2011), the data show that power relations of gendered identities and linguistic
practices may play a big role in Turkish-Dutch bilinguals’ language dominance. In obedience
to their husband’s authority, some Turkish women are isolated at home and have little
opportunity to practise Dutch. In the Turkish immigrant context, AoA accompanied by school
attendance in the Netherlands may provide better estimates of Dutch exposure and
proficiency. Likewise, language use and practices along with social factors (i.e. neighborhood
engagements, family members’ proficiency) seem to have a strong influence on Turkish-

26
Dutch bilinguals’ language knowledge.
Language choices of Turkish-Dutch bilinguals vary particularly in the domains of home
and media/leisure time activities. In addition to the differences in language preferences
across generations, wide variation also exists within the same generation, even within the
same family, as in the case where younger siblings are exposed to Dutch more often and
earlier than older ones. Nevertheless, potential variation within the family was not tested
systematically with the current methodology and data. Further studies should examine the
intra-family variation more closely. Another important phenomenon in this community is
that particularly among second- and third-generation bilinguals, using Dutch words while
speaking Turkish (i.e. code-switching) has become a common language style. According to
some bilinguals, this frequent mixing of Turkish and Dutch arises from their limited Turkish
vocabulary and, along with other factors (i.e. excessive exposure to Dutch, influence of
Dutch), negatively influences their Turkish proficiency. Therefore, the Turkish immigrant
community in the Netherlands may no longer be as linguistically homogeneous as once
observed. The dissolution of homogeneity can be a sign of social change in which
maintaining the Turkish language has become a challenge, whereas speaking Dutch is a
necessity in the Netherlands.
The interviews with Turkish-Dutch bilinguals regarding their attitudes towards
languages, motivations for learning languages, opinions about language choice, and their
experiences also pointed to a sign of ongoing language shift. Bilinguals in all three
generations emphasize their concerns about the decrease in Turkish use and knowledge of
third-generation bilinguals. They stress that the Turkish they speak in the Netherlands differs
from the Turkish spoken in Turkey. All of them exhibit very positive attitudes towards both
Turkish and Dutch but with a range of different motives. For first-generation bilinguals,
Turkish is of greater value than Dutch for reasons of identity, culture and religion. Although
the majority of second- and third-generation bilinguals also display positive attitudes that
reveal a deep commitment to Turkish culture and identity, they also express an emotional
attachment to the Netherlands as their birth country. They acknowledge the increasing
importance of knowing Dutch as opposed to Turkish, and the power this yields. Most feel
that their economic security and future prospects are largely dependent on their eventual
attainment level in Dutch. Additionally, the interview data unveil the perceived sociolinguistic
need to shift to Dutch, as well as parents’ expectations for their children’s academic

27
achievements, which may cause immigrant children to experience tensions and ambiguities
in the process of language maintenance and shift (Portes and Rumbaut 2001; Canagarajah
2008).
Furthermore, the research findings show that the relationship between Turkish identity
and language maintenance may not be unidirectional. As Edwards (1984) argues, a continuity
of identity may not necessarily be dependent upon communicative language retention.
Despite Turkish immigrants’ commitment to retaining Turkish language and identity, the
decline observed in third-generation bilinguals’ proficiency and use of Turkish confirms that
language maintenance fades among third-generation bilinguals. Although they value Turkish,
the findings on the language use and practices indicate that many third-generation bilinguals
make no specific efforts in practice to maintain it. Many prefer Dutch when responding to
their parents since they feel more comfortable speaking Dutch than Turkish. This situation
seems to cause socioemotional pressure concerning Turkish language maintenance,
triggering intergenerational tension within the family.
The nexus of Turkish language and identity among the young generation, however,
cannot be precisely described from the current study alone. Further research is necessary to
investigate the intricate connections between identity and language shift. In addition, many
other compelling factors related to Turkish immigrants’ experiences both in their home and
their host countries remain unexplored. For instance, the socioemotional consequences of
self-perceived language shift of these community members for which ethnic identity is a
core value (Extra and Yağmur 2004) would be worth considering for further exploration.
Linguistic anxiety of immigrants, as one of the possible effects of the pressure and tension
that immigrants experience within and outside the family would also be a valuable subject
for future research. Note, again, that all the results of this study must be interpreted with
caution, since they are based on a limited number of participants particularly in the younger
generation (n=41). For a thorough examination of linguistic and social components of
language shift, further studies are needed that include more participants from different
social backgrounds and combine different methods (e.g. linguistic analysis, self-reports or
interviews).

28
Acknowledgements
The author thanks Ad Backus, Marianne Gullberg, Anne Golden, her colleagues at MultiLing,
and the anonymous reviewers for their helpful comments and suggestions on earlier version
of this article.

Funding
This study was financially supported by the Research Council of Norway through its Centres
of Excellence funding scheme, project number 223265.

Appendices
Appendix A
Questionnaire
Appendix B
Interview Questions

References
Alders, Maarten, & Ingeborg Keij. 2001. Statistics Netherlands. Retrieved 4 27, 2014, from
third generation Turkish and Moroccan immigrants still small: http://www.cbs.nl/en-
GB/menu/themas/bevolking/publicaties/artikelen/archief/2001/2001-0813-wm.htm.
Backus, Ad. 2013. Turkish as an immigrant language in Europe. In Tej K. Bhatia and William C.
Ritchie (eds.), The Handbook of Bilingualism and Multilingualism, 2nd edn., 770–790.
Malden, MA: Wiley Blackwell.
Baker, Colin. 2000. A parents’ and teachers’ guide to bilingualism, 2nd edn. Clevedon:
Multilingual Matters.
Blom, Elma. 2010. Effects of input on the early grammatical development of bilingual
children. International Journal of Bilingualism 14. 422–446.
Bybee, Joan. 2010. Language, usage and cognition. Cambridge: Cambridge University Press.
Canagarajah, Suresh A. 2008. Language shift and the family: Questions from the Sri Lankan
Tamil diaspora. Journal of Sociolinguistics 12(2). 143–176.
Cohen, Jacob. 1992. A power primer. Psychological Bulletin, 112(1). 155–159.
Creswell, John W. & Vicki L. Plano Clark. 2011. Designing and conducting mixed methods
research, 2nd edn. Thousand Oaks, CA: Sage.

29
Dewaele, Jean-Marc. 2004. Perceived language dominance and language preference for
emotional speech: The implications for attrition research. In Monika S. Schmid,
Barbara Köpke, Merel Kejser, & Lina Weilemar (eds.), First language attrition:
Interdisciplinary perspectives on methodological issues, 81–104.
Amsterdam/Philadelphia: John Benjamins.
Dewaele, Jean-Marc & Aneta Pavlenko. 2001-2003. Web questionnaire on bilingualism and
emotion. Unpublished manuscript, University of London.
DeKeyser, Robert, Iris Alfi-Shabtay & Dorit Ravid. 2010. Cross-linguistic evidence for the
nature of age effects in second language acquisition. Applied Psycholinguistics 31. 413–
438.
De Klerk, Vivian. 2000. Language shift in Grahamstown: as case study of selected Xhosa-
speakers. International Journal of the Sociology of Language 146. 87–110.
Doğruöz, A. Seza & Ad Backus. 2009. Innovative constructions in Dutch Turkish: An
assessment of on-going contact-induced change, Bilingualism: Language and
Cognition, 12 (1). 41–63.
Doğruöz, A. Seza, & Stefan Th. Gries. 2012. Spread of on-going changes in an immigrant
language. Review of Cognitive Linguistics 10(2). 401–426.
Driessen, Geert, Frans Van der Slik & Kees de Bot. 2002. Home language and language
proficiency: A Large-scale Longitudinal Study in Dutch Primary Schools. Journal of
Multilingual and Multicultural Development 23(3). 175–194.
Ellis, Rod. 1994. The Study of Second Language Acquisition. Oxford: Oxford University Press.
Edwards, John. 1984. Language, diversity and identity. In John Edwards (ed.), Linguistic
Minorities, Policies and Pluralism, 277-310. London: Academic Press.
Extra, Guus & Kutlay Yağmur. 2004. Urban multilingualism in Europe: Immigrant minority
languages at home and school. Clevedon: Multilingual Matters.
Extra, Guus & Kutlay Yağmur. 2010. Language proficiency and socio-cultural orientation of
Turkish and Moroccan youngsters in the Netherlands. Language and Education 24(2).
117–132.
Field, Andy. 2013. Discovering statistics using IBM SPSS Statistics, 4th edn. London: Sage
Fishman, Joshua A. 1972. The sociology of language. Rowley, MA: Newbury House.
Gal, Susan. 1979. Language shift: Social determinants of language change in bilingual
Austria. New York: Academic Press.

30
Goulbourne, Harry, Tracey Reynolds, John Solomos, & Elisabetta Zontini. 2010. Transnational
families. Ethnicities, identities and social capital. London: Routledge.
Grenoble, Lenore A. & Lindsay J. Whaley. 2006. Saving languages. An introduction to
language revitalization. Cambridge: Cambridge University Press.
Grosjean, Francois. 1994. Individual Bilingualism. In Robert E. Asher & Jimmy M. Y. Simpson
(eds.), The Encyclopedia of Language and Linguistics, volume 3, 1656–1660. Oxford:
Pergamon Press.
Hlavac, Jim. 2013. Multilinguals and their sociolinguistic profiles: observations on language
use amongst three vintages of migrants in Melbourne. International Journal of
Multilingualism 10(4). 411–440.
Hulsen, Madeleine, Kees de Bot & Bert Weltens. 2002. “Between two worlds.” Social
networks, language shift, and language processing in three generations of Dutch
migrants in New Zealand. International Journal of the Sociology of Language 153. 27–
52.
Jamai, Abdeslam. 2008. Language use and maintenance among the Moroccan minority in
Britain. PhD thesis. University of Salford. Salford: UK.
Kang, Hyun-Sook & In-Sop Kim. 2012. Perceived and actual competence and ethnic identity
in heritage language learning: A case of Korean-American college students.
International Journal of Education and Bilingualism 15(3). 279–294.
Labov, William. 1972. Sociolinguistic patterns. Philadelphia: University of Pennsylvania Press.
Lanza, Elizabeth & Bente A. Svendsen. 2007. Tell me who your friends are and I might be
able to tell you what language(s) you speak: Social network analysis, multilingualism,
and identity. International Journal of Bilingualism 11(3). 275–300.
Leseman, Paul P. M. 2000. Bilingual vocabulary development of Turkish preschoolers in the
Netherlands. Journal of Multilingual and Multicultural Development 21(2). 93–112.
Li, Wei. 1994. Three generations, two languages, one family: Language choice and language
shift in a Chinese community in Britain. Clevedon: Multilingual Matters.
MacLeod, Andrea A. N. & Carol Stoel-Gammon. 2010. What is the impact of age of second
language acquisition on the production of consonants and vowels among childhood
bilinguals? International Journal of Bilingualism 14(4). 400–421.

31
Milewski, Nadja & Christelle Hamel. 2010. Union formation and partner choice in a
transnational context: The case of descendants of Turkish immigrants in France.
International Migration Review 44(3). 615–658.
Milroy, Lesley. 1980. Language and social networks. Oxford: Blackwell.
Montrul, Silvina. 2013. Bilingualism and the heritage language speaker. In Tej K. Bhatia &
William C. Ritchie (eds.), The Handbook of Bilingualism and Multilingualism, 2nd edn.,
168–189. Malden, MA: Wiley Blackwell.
Oh, Janet S. & Fuligni Andrew J. 2010. The role of heritage language development in the
ethnic identity and family relationships of adolescents from immigrant backgrounds.
Social Development 19(1). 202–220.
Pavlenko, Aneta. 2014. The Bilingual Mind: And what it tells us about language and thought.
Cambridge: Cambridge University Press.
Piller, Ingrid & Aneta Pavlenko. 2009. Globalization, multilingualism, and gender: Looking
into the future. In Li Wei & Vivian Cook (eds.), Contemporary Applied Linguistics 2:
Linguistics for the real world, 10–27. London: Continuum.
Romaine, Suzanne. 1995. Bilingualism, 2nd edn. Oxford: Blackwell.
Portes, Alejandro & Ruben G. Rumbaut. 2001. Legacies: The story of the immigrant second
generation. Berkeley, CA: University of California Press and Russell Sage Foundation.
Portes, Alejandro & Lingxin Hao. 2002. The price of uniformity: Language, family and
personality adjustment in the immigrant second generation. Ethnic and racial studies
25. 889–912.
Saltarelli, Mario & Susan Gonzo. 1977. Migrant languages: Linguistic change in progress. In
Marcel De Grève and Eddy Rosseel (eds.), Problèmes Linguistiques des Enfants des
Travailleurs Migrants, 167-186. Bruxelles: Didier.
Schleef, Erik. 2013. Written surveys and questionnaires. In Janet Holmes & Kirk Hazen (eds.),
Research Methods in Sociolinguistics, 42-57. Oxford: Wiley Blackwell.
Sevinç, Yeşim. 2014. Linguistic and social factors in Turkish-Dutch contact across generations.
Dutch Journal of Applied Linguistics 3(1). 82–100.
Sharma, Devyani. 2011. Style repertoire and social change in British Asian English. Journal of
Sociolinguistics 15(4). 464–492.
Silva-Corvalán, Carmen. 1994. Language Contact and Change: Spanish in Los Angeles. Oxford:
Clarendon Press.

32
Stevens, Gillian. 1999. Age at immigration and second language proficiency among foreign-
born adults. Language in society 28. 555–578.
Stoessel, Saskia. 2002. Investigating the role of social networks in language maintenance and
shift. International Journal of the Sociology of Language 153. 93–131.
Strauss, Anselm & Juliet Corbin. 1990. Basics of qualitative research. Newbury Park, CA: Sage.
Thomason, Sarah G. 2008. Social and linguistic factors as predictors of contact-induced
change. Journal of Language Contact 2(1). 42–56.
Valdés, Guadalupe, Joshua A. Fishman, Rebecca Chávez & William Pérez. 2006. Developing
minority language resources. The case of Spanish in California. Clevedon: Multilingual
Matters.
Vedder, Paul & Virta, Erkki. 2005. Language, ethnic identity and the adaptation of Turkish
immigrant youth in the Netherlands and Sweden. International Journal of Intercultural
Relations, 29(3). 317-337.
Yağmur, Kutlay. 2009. Language use and ethnolinguistic vitality of Turkish compared with the
Dutch in the Netherlands. Journal of Multilingual and Multicultural Development 30(3).
219–233.
Yılmaz, Gülsen & Monika S. Schmid. 2015. Second language development in a migrant
context: Turkish community in the Netherlands. International Journal of the Sociology
of Language 236. 101–132.

Author biography
Yeşim Sevinç is a doctoral student at University of Oslo, Center for Multilingualism in Society
Across the Lifespan. She received her BA degree in TESL at Istanbul University and her first
MA degree in Linguistics at Kafkas University. She obtained her second MA degree in
Linguistics at Radboud University Nijmegen. In her first MA thesis, she studied individual
differences between Polish and Turkish ESL learners. With her second MA thesis on the
Turkish immigrant community in the Netherlands, she won the third best MA thesis prize in
Belgium and the Netherlands awarded by Anela/VIOT in 2012. Her research interests include
linguistic, social, psychological and physiological aspects of multilingualism with a particular
focus on the immigrant context.

33